ARTICLE IN PRESS
JGLR-00130; No. of pages: 11; 4C: 4, 5
Journal of Great Lakes Research xxx (2010) xxx–xxx
Contents lists available at ScienceDirect
Journal of Great Lakes Research
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / j g l r
In situ-measured primary production in Lake Superior
Robert W. Sterner ⁎
Department of Ecology, Evolution and Behavior, University of Minnesota, 1987 Upper Buford Circle, St. Paul, MN 55108, USA
a r t i c l e
i n f o
Article history:
Received 1 June 2009
Accepted 19 November 2009
Available online xxxx
Communicated by Marlene Evans
Index words:
Carbon
Chlorophyll
DCM
Light
Temperature
a b s t r a c t
Water column primary production is a major term in the organic carbon cycle, particularly in large lakes with
relatively reduced shoreline and near-shore influence. Presently, there is a large imbalance in the known
inputs vs. outputs of organic carbon in Lake Superior. This study examined primary production in offshore
Lake Superior using in situ incubations over a range of conditions representing an annual cycle. Primary
producers were dominated by small (b 20 μm) cells and included a relatively large abundance of small,
spherical flagellates. During conditions with a warm surface layer, chlorophyll concentrations were two- to
three-fold higher within the deep chlorophyll maximum (DCM) than at the surface. Volumetric production
(mass L− 1 d− 1) was maximal at 2-10 m depth, well above the typical DCM depth. On average, 22% of 14C
label appeared in the dissolved pool at the end of the incubation period with the rest appearing in GF/Fstrained particles. A statistical model for volumetric production explained 93% of the variance in individual
measurements for depths N 2 m, using temperature and light as predictors. This model was applied to annual
fields of temperature and light, and a new estimate for whole-lake annual primary production, 9.73 Tg y− 1,
was derived. This combination of new measurements and modeling results brings the organic carbon cycle of
Lake Superior closer to being balanced.
© 2009 Elsevier B.V. All rights reserved.
Introduction
A well-constrained organic carbon budget is foundational to
understanding many aspects of the dynamics and biogeochemistry of
any given ecosystem. As presently understood, Lake Superior's organic
carbon budget is highly imbalanced. Cotner et al. (2004) estimated the
inputs to be as follows: atmospheric deposition, 0.16–0.41, rivers, 0.54–
0.62, and primary production, 5.3–8.2 (all in Tg y− 1). They estimated
losses to be as follows: outflow, 0.08–0.1, respiration, 13–39, and burial,
0.48 (all in Tg y− 1). Urban et al. (2005) estimated inputs to be:
precipitation, 0.1, rivers, 0.9, and production, 3–8 (all in Tg y− 1). They
estimated losses to be as follows: outflow, 0.1, respiration, 13–81, and
burial, 0.5 (all in Tg y− 1). According to these two budgets, respiration
dominates all other fluxes and organic carbon losses exceed gains by as
little as 1.5-fold or as much as 20-fold. The lack of balance in the organic
carbon budget signals a significant gap in the present understanding of
the dynamics of the Lake Superior ecosystem. Unless the lake's organic
carbon budget is not in equilibrium, one or more terms needs to be
revised (Cotner et al. 2004; Urban et al. 2005).
Primary production is thought to be the major source of organic
carbon in Lake Superior. Over the past several decades, there have
been only a few studies where primary production in Lake Superior
was measured. Examination of the methodological details these
⁎ Tel.: +1 612 625 6790.
E-mail address: [email protected].
studies used raises questions about relevance, representativeness, and
appropriateness of the numbers that have been used to construct
lake-wide, annual figures. The first attempt, a report by Putnam and
Olson (1961), describes work conducted during 1960 and indicated
that the O2 evolution method was inadequate to measure production
in this unproductive lake. These investigators switched to using the
14
C method and, in Putnam and Olson (1966), they reported the first
values for primary production in Lake Superior. Their studies were
performed on seven dates from July to September 1961 at one nearshore but deep water station in the western arm. Incubations were
performed in situ for 6 to 9 h beginning at solar noon. Six depths to
20 m were studied. Values ranged from 0.6 to 42.8 mg C m− 3 d− 1 and
from 76 to 507 mg C m− 2 d− 1. Later, Olson and Odlaug (1966)
returned to this “section of the lake” and measured primary
production in nine depth profiles to 20 m during July and August
1964; data are presented graphically in somewhat stylized fashion
making them difficult to use for any careful comparisons. Their
information indicates a volumetric production of 1.8 to 8.4 mg
C m− 3 d− 1 and a maximum areal production of 183 mg C m− 2 d− 1.
Parkos et al. (1969) reported on results of 201 measurements from
nine E–W transects performed on commercial shipping vessels. Water
collected from the surface was incubated at a single light level at one
temperature. There are some errors in the data presentation, whether
units are mg C m− 3 or mg C L− 1, for instance, but with some
guesswork a range of volumetric production of 0.3 to 38.7 mg
C m− 3 d− 1 (mean = 16.6, one very high outlier near the Duluth
harbor omitted) can be reconstructed. Vollenweider et al. (1974)
0380-1330/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jglr.2009.12.007
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
2
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
reported on a series of measurements performed during six cruises
April–December 1973 based on incubator conditions at nearoptimum light levels, “supplemented in each cruise by at least one
in situ experiment” (no further details are given), and they used a
model to integrate over the water column. Areal production was 330–
350 mg C m− 2 d− 1. Nalewajko et al. (1981) studied three sites during
summer, 1979. Water was collected from the epilimnion and
incubated at constant temperature. Mean volumetric production at
optimum light was 63 mg C m− 3 d− 1. Fahnenstiel and Glime (1983)
measured in situ production ten times from May through October
1979 at a deep water station 25 km from shore. Incubations ran for 4 h
during mid-day. Volumetric production ranged from b0.5 to N70 mg
C m− 3 d− 1 with a prominent subsurface maximum, particularly in
September and October. Nalewajko and Voltolina (1986) reported on
data collected from 37 stations during six cruises from May 1980 to
October 1981. Incubations were done in a light gradient at ambient
temperature. Volumetric rates at optimal light averaged 24 mg
C m− 3 d− 1 during stratified and 11 mg C m− 3 d− 1 during unstratified
conditions.
Fee et al. (1992) studied primary production in Lake Superior at a
single station near Isle Royal during 1990 and 1991. Reported rates
represent June 1 to August 31. Areal rates were calculated from P vs. I
curves, PAR extinction with depth, the level of PAR at the surface of
the lake, and the depth of the mixed layer (based on temperature
profile). Extracting data from their Figure 11 and converting units
indicates a range of volumetric production of 11–22 mg C m− 3 d− 1
and a range of areal production of 100–200 mg C m− 2 d− 1. Urban
et al. (2005) studied a series of sites mostly within 20 km of the
Keweenaw peninsula from May to July in 1998 and May to October in
1999. Most incubations were done at saturating light and application
to the water column was made via eight P vs. I curves. Areal
production was between b50 and 200 mg C m− 2 d− 1. Although the
above studies provide critical information on primary production in
the lake, there are multiple shortcomings involved in scaling these
observations up to an annual lake-wide value, as is needed to evaluate
the organic carbon budget.
From the above it is apparent that there have been very few
measurements performed offshore and fewer still done with in situ
incubations. Measurements done in the lab or lake for less than one
daylight period have to be extrapolated to an entire daylight period.
Perhaps most critically, the stratified season represents only ∼40–50%
of the year (Austin and Colman 2008), and observations of production
outside of that period are very rare. Finally, it is now understood that
measurements of primary production are potentially subject to a
variety of errors associated with contamination from trace metals or
deleterious effects of rubber associated with sampling and incubations, which can substantially reduce observed rates (Fahnenstiel et
al. (2002) provide a Great Lakes example). For all of these reasons, it is
natural to question whether the imbalanced organic carbon cycle in
Lake Superior results from inaccurate estimates of annual primary
production.
Nearly all the above studies considered depths b 20 m but during
the stratified season there is a deep chlorophyll maximum below this
depth; thus, a large proportion of the photosynthetic machinery has
been left out of prior estimates. The highest concentration of
chlorophyll in Lake Superior typically occurs below the surface
mixed layer during the stratified season (Moll and Stoermer 1982;
Fahnenstiel and Glime 1983; Barbiero and Tuchman 2001). Repeated
offshore sampling over the past ∼ 10 y (Sterner, unpublished)
indicates that a deep chlorophyll maximum (DCM) is a persistent
and predictable feature of the entire stratified season. Given the
prominence of this feature of the photosynthetic biomass of Lake
Superior, the activity of the phytoplankton within the DCM might play
a large role in production dynamics of the lake. Lake Superior's DCM is
thought to represent both an increased chlorophyll content of shadeadapted algae and a higher level of biomass (Barbiero and Tuchman
2001). The mechanisms behind the formation and maintenance of the
DCM in Lake Superior are still not well worked out (Barbiero and
Tuchman 2001). Relationships between CTD-measured fluorescence,
chlorophyll measured on extractions from filtered samples, and
biomass as determined microscopically or as particular organic
carbon are complex and vary across the five Laurentian Great Lakes
(Barbiero and Tuchman 2001). Fahnenstiel and Glime (1983)
presented the most detailed information available to date on rates
of production relative to the DCM in Lake Superior. They observed a
correspondence between the development of the DCM and subsurface
levels of primary production and concluded that in situ production
was the principal process associated with the development of the
DCM.
In terms of nutrients, Lake Superior algae are mainly P limited, with
Fe deficiencies also noted when P is supplied, indicating that primary
producers are near “the cusp” of macro- and micronutrient limitation
(Sterner et al. 2004). Algal nutrient status (for example, as measured
by C:P ratios) indicate greater nutrient content of algae at the DCM and
below than in surface waters (Barbiero and Tuchman 2001). However,
in spite of several careful examinations of distributions of very low
dissolved phosphate in the lake (Baehr and McManus 2003; Field and
Sherrell 2003; Anagnostou and Sherrell 2008), there is no clear
evidence for nutrient gradients with depth that would support
upwelling as a primary mechanism for DCM formation. Ammonium
does show significant vertical structure offshore late in the season
(Kumar et al. 2007), but it is highest at depths near the DCM rather
than below it. Lower seston C:P at depth may reflect reduced
photosynthesis (lowering C) as much as it may represent enhanced
availability of nutrients (increasing P), so nutrient supply as a
mechanism for DCM formation lacks unequivocal support. Loss rates
as a function of depth have not yet been reported and losses are a
potential mechanism for DCM formation and maintenance.
Detailed and comprehensive floristic study of Lake Superior
indicated an importance of phytoflagellates (cryptophytes, chrysophytes, and dinophytes) and diatoms with small contributions by
bluegreens and greens (Munawar and Munawar 1978). After
methodology for studying picoplankton became available it soon
became clear that this small size fraction was of great importance in
this lake. Fahnenstiel et al. (1986) estimated that approximately 50%
of autotrophic production of Lake Superior was attributable to
phytoplankton that passed through a 3 μm screen. Large contribution
of picoplankton production to the total was also noted by Ivanikova et
al. (2007). Lake Superior harbors unusual and perhaps unique
picocyanobacterial clades. Analysis of 16S rRNA genes of the Lake
Superior autotrophic picocyanobacteria indicates that the majority of
sequences clustered within the Synechococcus and Prochlorococcus
clade, and that the most abundant species in the offshore are two new
clusters of Synechococcus, which so far have not been described in any
other environment (Ivanikova et al. 2007; Ivanikova et al. 2008). In
contrast, picocyanobacteria in a Lake Superior bay corresponded to
the cosmopolitan Synechococcus that has been found in numerous
locations worldwide (Ivanikova et al. 2007).
In this study, primary production at offshore stations in Lake
Superior over a range of stratified and unstratified conditions was
measured using a standardized in situ protocol. Measurements were
made to either 80 or 100 m depth, i.e. above, within and below the
DCM. Precautions to minimize trace metal contamination were taken.
Production was routinely measured as the incorporation of tracer into
particles sampled on glass fiber filters at the end of the incubation.
Because of the possibility that some 14C tracer may cycle into the
dissolved pool in b24 h, the total organic carbon incorporation
including DOC as well as particles was measured in a subset of cruises.
A predictive model for production as a function of light and
temperature was used to scale “snapshot” measurements to an entire
annual cycle. Although logistical difficulties still limited the manageable spatial and temporal coverage, the use of day-long, in situ
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
incubations, offshore locations, extension through and below the
DCM, and inclusion of measurements outside of the warm stratified
period make this study unique compared to other studies of primary
production in Lake Superior.
Methods
The two earliest deployments (one “deployment” is a single
incubation of a cabled array of bottles incubated at eight different
depths) occurred at sites CD-1 (47°3′54″N, 91°25′55″W) and EM
(47°34′58.8″, 88°51′0″) (for a map, see Kumar et al. 2008). Beginning
in July 2007, deployments were performed at station WM (47°20′0″,
89°48′0″). Temperature and other physical and biological parameters
were measured via CTD (Supplementary Table 1). Water for
incubations and for additional chemical analysis was collected using
Niskin bottles modified to eliminate all rubber, latex, or metal in the
internal parts. The central bungee was high-purity silicon pump
tubing (food grade) and O-rings also were silicon.
Particulate nutrients were routinely prefiltered with 80-μm Nitex
to eliminate the occasional large grazer from the sample. During the
period of this study, algal biomass retained within routine 80-μm
zooplankton nets was not visible, indicating that prefiltering did not
remove a significant pool of photosynthetically active biomass from
the present measurements. Beginning in July 2007, seston was
measured in size fractions b2 μm, b10 μm, and b80 μm and
chlorophyll was measured in size fractions b2 μm, b5 μm, b10 μm,
b20 μm, b40 μm, and b80 μm. Particulate organic carbon (POC),
particulate organic nitrogen (PON), and particulate phosphorus (PP)
samples were collected by filtering 1 to 1.5 L of sample water onto a
0.7-μm GF/F filter (preashed at 450 °C for 4 h to remove residual
carbon, PP filters then rinsed with 5 mL of 1% HCl and 50 mL of
Millipore water). Samples were transported frozen then dried at 60 °C
for 24 to 48 h and prepared for analysis. POC and PON samples were
analyzed on a Perkin Elmer 2400 CHN analyzer calibrated with an
acetanilide standard. PP samples were digested with potassium
persulfate and analyzed using the ascorbic acid method. DIC and
DOC samples were collected in preashed brown glass vials, prerinsed
with sample water then filled to overflowing and capped with a
septum and cap. Samples were analyzed within 4 days of collection on
a Shimadzu TOC-Vcsh analyzer using a platinum catalyst.
Nitrate concentrations at numerous depths within the water
column were measured on multiple cruises during 2005–2008.
Unfiltered samples were collected in 1% acid cleaned HDPE bottles
and frozen for transport and storage. Samples were analyzed
using an OI Analytical FS 3000 automated chemistry analyzer
(2005–2007) or a Lachat QuikChem® 8500 Flow Injection Analysis
System (2007–2008) and a cadmium reduction column. Column
reduction efficiency was routinely checked by comparing nitrite
vs. nitrate recovery. These data were combined into a single
seasonal trend to calculate the rate of nitrate drawdown in the
mixed layer.
3
Phytoplankton samples were collected at all incubation depths
beginning in July 2007, preserved with Lugol's, and kept under cold
and dark conditions. Samples of 100 mL were settled in graduated
cylinders for at least 20 h and the overlying 90 mL was removed by
aspiration. These 10-mL concentrated samples were then placed in
counting chambers and allowed to settle overnight. Samples were
observed with an inverted microscope at 400× and cells N 3 μm in at
least one dimension were quantified. Randomly chosen grids were
examined until at least 75 total cells (or, for some deep samples with
very low algal density, 25 total cells) were enumerated. Taxa were
identified using several keys and with reference to the species list
compiled by Munawar and Munawar (1978). Much more intensive
methods than this would be required for high accuracy taxonomic
identifications, but this approach provides a general indication of the
species composition of the larger algae encountered in this study.
In situ measurements of primary production followed the JGOFS
protocols (Knap et al. 1996). The first deployment at CD-1 occurred
from 1115 to 2130 and rates were adjusted to the full day-length
period. For the remainder of deployments, samples were collected at
least 1.5 h before sunrise (Table 1) from incubation depths. Tracemetal-cleaned 250-mL polycarbonate Nalgene bottles were rinsed 3×
with lake water and filled to the brim. Four light and one dark bottle
were filled at each depth. These were shielded from deck work lights
and transferred into a radiation safety van where work continued
under low levels of red light. Bottles were opened and spiked with
0.25 mL of NaH14CO3 (∼20 μCi, MP Biomedicals no. 17441H25, specific
activity 30–60 mCi/mmol, not chelex rinsed). At each depth, the
entire contents of one light bottle was immediately filtered onto
precombusted 25 mm GF/F filters to estimate T0 14C uptake. Bottles
were then capped, sealed with Parafilm® and cable-tied into wire
baskets. These baskets were then suspended from a cable hanging
from a float. A secondary float was tied to the main float and was
equipped with a strobe light, radar reflector and PAR sensor (Onset SLIA-M003) which measured light between 0° and 80° from vertical
when pointing straight upwards.
PAR data was collected at 1-minute intervals from approximately
1 h before sunrise to approximately 1 h after sunset. The array of
bottles drifted freely during the incubation period. At 1.5 h after
sunset or later, bottles were retrieved and moved again into the
radiation safety van where work continued under low levels of red
light. Each bottle was opened, and 0.25 mL was removed and
transferred into a scintillation vial containing 0.25 mL β-phenylethylamine and scintillation fluor to measure total radioactivity. In a
subset of cruises, a second aliquot of 1 mL was removed and
transferred into a vial with 250 μL of 6N HCl. These were shaken for
N1.5 h to remove inorganic 14CO2 and are referred to here as “total
organic carbon” measurements. The remaining contents, or for some
depths with high particle density, subsamples of the remaining
contents, were then filtered onto 25-mm precombusted GF/F filters
and transferred to scintillation vials. Vials containing filters were later
acidified with 0.25 mL 0.5N HCl, evaporated to dryness, filled with
Table 1
Light conditions. Daily integrated PAR as measured on drifter array buoy (beginning in 2007). Apparent sunrise and sunset as calculated from http://www.srrb.noaa.gov/highlights/
sunrise/sunrise.html/ without adjusting for daylight savings time. Extinction coefficient (K) taken from CTD profiles of PAR, eliminating data from near surface and depth to produce
strong linear plot of log PAR vs. depth. Thermocline depth (T, m) was calculated by picking the midpoint of the two 1-m binned measurements of temperature associated with the
greatest temperature gradient.
Date
PAR (photons m− 2 d− 1 × 1025)
Sunrise
Sunset
Conditions
K (m− 1)
Z1%
June 19, 2006
August 7, 2006
July 31, 2007
November 7, 2007
April 30, 2008
July 31, 2008
September 17, 2008
n.d.
n.d.
2.50
0.138
2.41
2.71
2.16
0410
0441
0436
0653
0444
0437
0540
2004
1921
1934
1633
1910
1933
1806
Clear, sunny
Clear, sunny
Hazy sunshine
Cloudy throughout
Hazy sunshine
Partly cloudy
Clear, sunny
0.126
0.115
0.126
0.114
0.117
0.127
0.161
36.5
40.0
36.5
40.4
39.4
36.3
28.6
a
light
(m)
T (m)
13.5
18.5
7.5
66.5
127.5a
8.5
16.5
Early-season, reverse stratification.
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
4
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
fluor and counted on a Beckman-Coulter LS 6500 liquid scintillation
counter within 10 days of collection. Primary production using filter
samples was calculated as:
−1 −1
Particulate Production mg C L d
= ðF = V Þ ð0:00025C = T Þ
ð1:05 = DÞ
ð1Þ
For deep incubations, T0 values were larger, sometimes 50% of sample
DPMs.
Primary production from total organic carbon samples was
calculated as
−1 −1
= ðP = 0:001Þ
Total Organic Production mg C L d
ð0:00025C = T Þ ð1:05 = DÞ
where:
F = DPMs in filtered sample, corrected for T0
V = volume filtered (L)
C = DIC concentration (mg C L− 1)
T = total 14C DPMs (in 0.25 mL)
D = incubation duration (=1 day),
and with 0.00025 converting total DPMs to 1-L volume and 1.05
representing the correction for the lower uptake of 14C compared to
12
C. Production rates were corrected for both T0 and dark uptake. In
general, T0 DPMs were ≪1% of light sample DPMs for depth b20 m.
ð2Þ
where:
P = DPMs in 1-mL sample,
and with 0.001 converting mL to L.
OC samples were taken in light and dark bottles and were darkcorrected before comparison with particulate production numbers.
Column-integrated productivity was calculated using trapezoidal
integration and extrapolating the rate measured for the shallowest
depth to the surface. Volumetric production below the deepest
measurement was assumed to be zero.
Fig. 1. Vertical profiles of temperature, chlorophyll, and primary production for cruises during warm mixed layer conditions.
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
Results
Surface temperatures ranged from 1.8 to 18 °C, nearly the full
range ever observed in the lake. During summer stratification, all
variables associated with phytoplankton biomass (POC, PON, PP, and
chl) normally exhibited mid-depth maxima (see Fig. 1 and Supplementary Fig. 1). Additional physical and chemical parameters
measured at each incubation depth are given in Supplementary
Table 1. The chl:C ratio under stratified conditions was typically
greatest between 20 and 40 m depth (Supplementary Table 1). DOC
levels were somewhat higher in November and April than under
warmer conditions. Nitrate was always much higher in concentration
than ammonium. Stoichiometric ratios of C:N and C:P were often
highest near the surface. Daily incident photon flux ranged more than
10-fold from the lowest to the highest observed value (Table 1). The
PAR extinction coefficient (K) on the other hand was similar from
deployment to deployment, so that the depth of 1% surface light
ranged only between 29 and 40 m (Table 1). The relatively small
variation in K is a result of year-round low levels of chlorophyll in the
lake. Thermocline depth varied greatly by deployment, from a
minimum of 7.5 m to a maximum of almost 130 m (Table 1).
The size structure of photosynthesizing organisms was strongly
skewed toward small organisms. Approximately half of the chlorophyll passed through a 2-μm filter, and approximately three quarters
5
of the chlorophyll passed through a 5-μm filter (Fig. 2A). POC size
fractions as a function of depth (Supplementary Fig. 1) also indicated
an importance of small particles. Under stratified conditions, at
depths b 50 m, more than half of the POC passed through a 2-μm filter.
At greater depth, the relative importance of this small size fraction
was typically even larger. Quantifiable algal taxa (Table 2) were
diverse and their abundances were low. Consistent with the detailed
floristic work done on the lake in the 1970s, small flagellates
(Ochromonas, Cryptomonas, Chroomonas, and others) were relatively
numerous. Previous authors have noted either a reduced abundance
of centric diatoms in the metalimnion (Barbiero and Tuchman 2001)
or an increased abundance of centric diatoms in the upper portion of
the DCM (Fahnenstiel and Glime 1983). Relatively high abundance of
centric diatoms was observed in the present study only on one date
(September 17, 2008, not shown).
Dark volumetric 14CO2 uptake for the first array deployment was
clearly higher than the others, which likely is because it alone was
begun during daylight and thus bottles received some illumination
during deployment (Table 3). In other deployments, dark volumetric
production was b1 mg C m− 3 d− 1, with greatest values observed in
the shallowest samples during stratified conditions (Table 3). Dark
corrections were made where indicated below but were comparable
to light production rates only at deepest incubation depth where
production rates were low; thus, whether dark bottles are subtracted
Fig. 2. (A) Chlorophyll size structure for all 2007 and 2008 cruises at all in situ incubation depths, given as the fraction of the chlorophyll measured in the largest size fraction
(b 80 μm) for that depth and date (N = 40 for all size fractions). Mean, SE (boxes), and SD (whiskers) all shown. (B) CO2 uptake into particles (N 0.07 μm) in light bottles for all in situ
incubations (mean of 3 bottles incubated at each depth). See also Fig. 1 and Table 3. (C) Residual plot obtained from regression modeling of individual measurements (depth N 2 m) of
production within in situ incubation bottles as a function of temperature and light (Eq. (3)). All cruises are included. November data are indicated by closed circles. (D) Modeled
water column production (mg C m− 2 d− 1) into particles as a function of day of year.
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
6
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
Table 2
Abundances of algal taxa observed at all incubation depths at site WM (5 dates times
8 depths = 40 samples). Percent occurrence refers to the percent of the 40 samples in
which that particular taxon was identified. “Taxon” refers to the most specific identity
known, which was used during enumeration.
Taxon
Cells/ml
SE
Percent occurrence
Bacillariophyta
Centric diatom
Synedra sp.
Asterionella formosa
Fragilaria sp.
Nitzschia sp.
Cymbella sp.
Melosira sp.
Navicula radiosa
39.0
30.1
4.5
2.3
0.8
0.4
0.2
0.2
8.0
4.7
2.4
1.2
0.8
0.4
0.2
0.2
90
76
14
10
2
2
2
2
Chlorophyta
Scenedesmus ellipticus
Tetraedron sp.
Ankistrodesmus falcatus
Scenedesmus sp.
42.8
4.2
2.3
0.3
13.5
1.4
1.0
0.3
55
24
17
2
Cryptophyta
Cryptomonas erosa
Chroomonas sp.
Cryptomonas ovata
37.0
30.4
1.2
5.0
8.1
0.6
83
62
12
Cyanobacteria
Oscillatoria limmetica
Microcystis aeruginosa
119.9
4.6
33.0
4.3
67
5
Ochrophyta
Dinobryon divergens
Rhizosolenia sp.
Dinobryon bavaricum
18.3
3.8
0.8
7.9
0.9
0.6
38
33
5
1.5
1.3
5
186.7
0.5
28.6
0.3
100
5
Dinophyta
Gymnodinium sp.
Unassigned
small spherical flagellates
Pseudokephyrion sp.
from light bottles or not plays little role in the areal production
estimates.
Light volumetric 14CO2 uptake profiles showed either an approximately exponential decline from surface downwards or a pronounced
surface inhibition (Figs. 1 and 2B). The three deployments where surface
inhibition was observed occurred during the early-mid portion of the
season (June 2006, July 2007, and April 2008), whereas those without
surface inhibition were performed in the mid-late portion of the season
(August 2008, November 2007, July 2008, and September 2008). At
depth ≥20 m, the five profiles studied during summer stratification
exhibited a high degree of consistency from profile to profile.
Production expressed as a ratio to chlorophyll or POC concentration
was always highest within the upper 10 m and usually was highest
within the upper 5 m (Table 3). In several deployments, production to
biomass (P/B) ratios exhibited reduced values in the shallowest depths.
Summertime maximum P/B ratios expressed per unit carbon were in
the range of 0.1 to 0.15 d− 1. However, P/B ratios at most depths were
considerably lower. During stratified conditions, nearly all deployments
showed highest P/B ratios at shallower depths than either the POC or
chlorophyll peaks. There was up to a 20-m separation between highest
production values (volumetric or P/B) and peak biomass. Neither the
DCM nor the biomass peak generally occurred at the depth of most
active production or biomass turnover gain due to primary production.
There are significant complications with respect to P/B ratios calculated
this way because POC includes non-autotrophic contribution. Given the
heavy picoplankton dominance of the phytoplankton in Lake Superior,
any attempt to calculate autotrophic biomass from routine microscopic
counts would be fraught with error so was not made.
Water column-integrated production was lowest in the November
deployment, higher in the April deployment, and highest and within a
narrow range for all deployments under warm stratified conditions
(Table 4). Total organic carbon production (dissolved plus particulate)
values were usually higher than particulate production values (Table
5). There was no clear relationship between the ratio of total to
particulate production as a function of depth or season. The average
value of this ratio was 1.28, meaning 0.28/1.28 = 22% of total carbon
fixed within a day appeared in the DOC pool at the end of that day.
To obtain an estimate for annual primary production from these
observations, a statistical model was created predicting the rates of
production measured in individual incubation bottles (n = 40) as a
function of temperature and light. Only cruises with incident PAR
measurements were included (5 cruises × 8 depths = 40 observations). Temperature came from CTD casts and daily integrated photon
flux came from combining information from daily incident flux taken
from the buoy with the extinction coefficient calculated from the CTD
casts. This predictive model was then applied to annual cycles of light
and temperature as a function of depth. A variety of temperature- and
light-dependent models from the literature were tried and evaluated
by examining r2 values and plots of residuals. Both Eppley (1972) and
Boltzmann (Gillooly et al. 2002) temperature dependence were
examined as was a variety of growth-irradiance curves of different
forms (Jassby and Platt 1976). Producing a successful model (r2 N 0.9
and no sign of bias in residual plots) required elimination of
observations from 2 m depth; these were sometimes high and
sometimes low and no model successfully predicted them. Usually,
more than 90% of measured water column production occurred at
depths N 2 m, so removing those depths from the predictive model did
not appear to be a serious shortcoming in calculating water column
production. The predictive model so chosen was:
ð−Ea = 0:00008624T Þ
P = Ce
−α I = Popt
Popt 1 − e
ð3Þ
where:
P = volumetric production (mg C m− 3 d− 1, light–dark)
C = constant (fit, =1159, dimensionless)
Ea = activation energy (fit, = 0.283, eV)
8.62 × 10− 5 = Boltzmann's constant (eV/K)
T = temperature (K)
Popt = a factor associated with the optimal rate of production at
high I (fit, =836, mg m− 3 d− 1) (note: this rate is not actually
achieved because other terms in the equation ≠ 1)
α = a parameter giving light dependence of production (fit, =7668,
m2 photons− 1 × 10− 25)
I = daily integral irradiance (photons m− 2 × 1025)
The model had good predictive power for summer and winter
samplings combined (r2 = 0.93) and was unbiased (Fig. 2C). Surprisingly, models that included chlorophyll concentration as a predictor
were always poorer from the standpoint of having lower r2 and in
terms of bias indicated in residual plots. The simpler model without
chlorophyll therefore was preferable for the present purposes.
Eq. (3) was then applied to matrices of temperature and light as a
function of day of year and depth. The temperature matrix was
produced by combining data from 65 offshore cruises undertaken by
the author's research group between 1996 and 2007. Dates spanned
January 31 to November 8 although most cruises occurred within
June–September. Observations of temperature were interpolated
using the software program SurGe (http://www.geocities.com/
miroslavdressler/surgemain.htm), which implements the ABOS approximation method (http://m.dressler.sweb.cz/ABOS.htm). Incident light as a function of month came from two sources, the
Vegetation/Ecosystem Modeling and Analysis Project (http://www.
cgd.ucar.edu/vemap/) and the National Renewable Energy Laboratory (http://www.nrel.gov/). Both sources provided modeled incident
light at points on the Lake Superior shoreline near the sampling sites.
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
7
Table 3
Volumetric production and productivity to biomass ratios. Values in bold signify the maximum observed value for that parameter for that date.
Date, Cruise
Depth (m)
June 19, 2006, NILS12
August 7, 2006, NILS14
July 31, 2007, CARGO1
November 8, 2007, CARGO3
April 30, 2008 , CARGO4
July 31, 2008, CARGO5
September 17, 2008, CARGO6
5
10
20
30
40
60
80
100
5
10
20
30
40
60
80
100
2
5
10
20
30
40
60
80
2
5
10
20
30
40
60
80
2
5
10
20
30
40
60
80
2
5
10
20
30
40
60
80
2
5
10
20
30
40
60
80
Production (P)
Biomass
POC b 80
P/chlorophyll
P/POC μmol C
Light
Dark
(μmol C L− 1 d− 1)
(μg chl L− 1)
(μmol C L− 1)
(μmol C μg chl− 1 d− 1)
(μmol C− 1 d− 1)
0.75
1.74
0.96
0.09
0.12
0.04
0.02
0.03
1.07
1.09
0.89
0.37
0.12
0.01
0.00
0.00
0.30
0.95
1.19
0.74
0.46
0.21
0.02
0.00
0.71
0.38
0.21
0.09
0.03
0.02
0.01
0.01
0.36
0.76
0.52
0.54
0.30
0.12
0.02
0.02
1.66
1.24
1.28
0.67
0.20
0.10
0.01
0.01
1.35
1.05
1.07
0.76
0.26
0.04
0.01
0.01
0.593
0.678
0.672
0.863
0.551
0.366
0.348
0.318
0.380
0.986
0.875
1.376
2.192
0.492
0.294
0.257
0.462
0.699
1.065
1.332
1.245
1.155
0.558
0.636
0.716
0.752
0.872
0.527
0.914
0.586
0.350
0.078
0.801
0.687
0.617
0.444
0.456
0.383
0.768
0.693
0.732
0.710
0.777
1.359
0.520
0.174
0.289
0.192
0.574
1.065
1.134
1.503
1.212
0.456
0.207
0.157
13.75
13.39
11.15
9.76
7.79
5.77
5.71
5.20
7.31
13.25
15.05
18.37
13.28
6.38
5.29
5.46
8.63
9.48
12.42
10.85
11.03
7.69
6.02
5.70
11.79
12.10
11.09
9.90
10.73
10.19
10.89
5.90
8.07
8.45
8.50
8.05
7.35
7.43
7.48
8.36
11.72
11.04
11.75
13.28
8.89
6.03
6.65
6.71
11.97
14.98
16.66
16.08
12.01
7.03
5.28
4.67
1.21
2.43
1.27
0.03
0.11
0.00
− 0.02
0.00
2.76
1.08
0.97
0.28
0.05
− 0.01
− 0.01
− 0.01
0.64
1.32
1.09
0.53
0.37
0.17
0.02
0.00
1.00
0.50
0.23
0.17
0.04
0.03
0.23
0.08
0.44
1.09
0.80
1.22
0.64
0.30
0.01
− 0.02
2.19
1.70
1.60
0.48
0.35
0.60
0.04
0.13
2.27
0.97
0.92
0.48
0.18
0.10
0.05
0.03
0.052
0.123
0.077
0.003
0.008
0.000
− 0.001
0.000
0.143
0.081
0.057
0.021
0.008
0.000
0.000
0.000
0.034
0.097
0.093
0.065
0.042
0.025
0.002
0.000
0.061
0.031
0.018
0.009
0.003
0.002
0.007
0.001
0.044
0.088
0.058
0.067
0.040
0.016
0.001
− 0.001
0.137
0.109
0.106
0.049
0.021
0.017
0.002
0.004
0.109
0.069
0.063
0.045
0.019
0.007
0.002
0.001
0.04
0.09
0.10
0.06
0.06
0.04
0.03
0.03
0.02
0.03
0.04
−0.02
0.01
0.01
0.01
0.01
lost
0.03
0.04
0.03
lost
0.02
0.00
0.00
0.00
0.01
0.01
0.00
− 0.01
0.00
−0.06
0.00
0.00
0.02
0.02
0.00
0.01
0.01
0.01
0.03
0.06
0.04
0.04
0.02
0.02
− 0.01
0.00
−0.01
0.05
0.02
0.03
0.04
0.04
− 0.01
0.00
0.00
These models were based on direct measurements at other stations
coupled to locally observed meteorological information on parameters such as cloudiness. These generated nearly identical measures
of production so only one is presented here. Incident solar radiation
from these sources (kJ m− 2 or kWh m− 2 d− 1, depending on the data
source) includes energy outside of the PAR band so radiation was
Table 4
Water column integrated primary production for each 1-d period. Tg y− 1 is given to enable ready comparison to past estimates of annual primary production rates.
Light
Dark
Light–dark
−2
−1
mg C m
d
Tg y− 1
−2 −1
mg C m
d
Tg y− 1
−2 −1
d
mg C m
Tg y− 1
June 19, 2006
August 7, 2006
July 31, 2007
November 8, 2007
April 30, 2008
July 31, 2008
September 17, 2008
383.4
11.5
62.3
1.9
321.2
9.62
369.3
11.1
17.2
0.52
352.1
10.6
351.9
10.6
15.2
0.46
336.7
10.1
88.6
2.7
1.2
0.04
87.4
2.62
227.4
6.8
11.2
0.34
216.2
6.48
372.6
11.2
12.6
0.38
360.0
10.8
336.3
10.1
13.7
0.41
322.6
9.67
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
8
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
Table 5
Total organic carbon production (TOCP) and the ratio of total (“T”) to particulate production (“P”).
Depth
2
5
10
20
30
November 2007
July 2008
September 2008
TOCP (mg C m− 3 d− 1)
T:P
TOCP (mg C m− 3 d− 1)
T:P
TOCP (mg C m− 3 d− 1)
T:P
8.94
4.27
–
1.36
0.76
1.0
0.9
–
1.3
2.4
24.82
20.09
18.83
9.66
3.63
1.2
1.3
1.2
1.2
1.5
29.94
14.65
14.61
10.00
2.30
1.8
1.2
1.1
1.1
0.7
converted to PAR although both unit conversion and empirically
adjusting the values for months when PAR from the incubation buoy
was available. The empirical adjustment was 0.42 for VEMAP and 0.51
for NREL (PAR ≈ 40–50% of total solar radiation). The seasonal
minimum PAR was estimated to be 5.88 × 1024 photons m− 2
(December) and the maximum seasonal PAR was estimate to be
3.09 × 1025 photons m− 2 (July). Because of minimal seasonal variation
in extinction coefficient observed in the present study, a single value
(0.13 m− 1) was used.
Water column-integrated production modeled in this way is given
in Fig. 2D. These calculations assume zero ice cover; we lack any
certain information on how to estimate primary production under ice.
Austin and Colman (2007; 2008) estimated that ice cover on Lake
Superior has decreased from 23% to 12% (spatially and temporally
averaged from Dec 1 to May 31) over the last century. According to
this model, production ranges from ∼ 200 to ∼350 mg C m− 2 d− 1.
Finally, to obtain a single annual estimate, the values in Fig. 2D were
summed with the earliest and latest values extrapolated to the first
and last days of the year, respectively. These earliest and latest
measured values are based on water column temperatures b 4 °C and
lowest or near-lowest monthly incident light and thus seem to be a
reasonable basis for extrapolation, considering the information on
incident light was monthly. The lake-wide, annual primary production value thus obtained was 93.7 g C m− 2 y− 1. Multiplying by the
surface area of the lake gives 7.6 Tg y− 1. These values, determined
with standard JGOFS protocols, are values that are most comparable to
other studies. Primary production channeled into particles is most
relevant to considerations of energy flow directly from phytoplankton
into upper trophic levels. However, consideration of the entire organic
carbon cycle should also consider primary production that cycles
quickly into the dissolved pool; this fraction is available for uptake by
heterotrophic activity. Multiplying by the average increase in
production due to 14C label that cycles directly into the DOC pool
within 1 day (1.28, Table 5), increases these annual estimates to 120 g
C m− 2 y− 1, or 9.73 Tg y− 1.
Nitrate concentration in the upper part of the water column was
lower than in the lower part of the water column, with an apparent
steady mixed layer drawdown (Fig. 3). The high concentration
observed in deep water on Oct 2, 2005 (site CD-1), seems anomalous
or may indicate a localized, late-season sediment efflux. The period
from April 29 to Aug 29 exhibits steady decline in nitrate
concentration in surface waters with a rate of depletion of 1.35 nM
−1
. On average, NO−
NO−
3 d
3 uptake represents 22.7% of summertime
total N uptake in Lake Superior (Table 1 in Kumar et al. 2008). Stable
isotope dynamics indicates less importance of nitrate toward total N
uptake in the winter (Kumar et al., submitted), although the
calculations performed here are based only on summer dynamics.
The average seston C:N ratios for depths ≤ 30 m during the period of
NO−
3 drawdown was 8.7 (molar, Supplementary Table 1). Combining
the above information, one can arrive at an estimate of C incorporation associated with the NO−
3 drawdown in the upper 30 m of the
water column:
−
1:35 nmol NO3 ðL dÞ
−1
−
1 N=0:227 NO3
−1
8:7C = N = 51:7 nmol C ðL dÞ
Fig. 3. (A) Seasonal NO−
3 concentration for different water column segments assembled
from several sampling years. Two deep water observations with concentrations
indicated by (+) were deemed to be outliers and were removed from the trends. (B)
Late-summer profiles of [NO−
3 ] and temperature in Lake Superior. Sampling locations
were: EM, 47°35′59″, 88°51′0″ and EL-1, 47°30′0″, 87°0′0″. Both profiles from late
August 2005 were chosen for display because they showed the greatest observed
difference between surface and deep water.
;
ð4Þ
or 129 mg C m− 2 d− 1. The carbon uptake associated with the seasonal
nitrate drawdown thus is less than half of the production measured by
the in situ 14C method (compare to values in Table 4). A term not
factored into the above calculation is any replenishment of the
epilimnetic nitrate pools either via mixing of sub-thermocline waters
into the epilimnion or water column nitrification. It is thus perhaps
best to consider this estimate of production based on nitrate
drawdown to be a minimum value. Reversing the chain of calculations, and estimating expected nitrate drawdown given the above
assumptions plus the observed primary productivity, we would
expect about twice the seasonal nitrate drawdown than what is
observed, possibly indicating a supply of nitrate to surface waters
from mixing or from nitrification or both approximately equal to the
−1 −1
d ).
drawdown rate (∼1.3 nmol NO−
3 L
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
Discussion
Recent reviews have placed whole-lake, annual primary production in Lake Superior at between 5.3 and 8.2 Tg y− 1 (Cotner et al.
2004) or between 3 and 8 Tg y− 1 (Urban et al. 2005) and have pointed
to a greatly imbalanced organic carbon budget, with outputs ≫ inputs.
The present study revises the annual primary production number
upwards to 9.73 Tg y− 1. Both Cotner et al. and Urban et al. estimated
the allochthonous supply of organic C to be ∼1 Tg y− 1. Thus, an
estimate for total annual gains of organic production in Lake Superior
is ∼ 10 Tg y− 1. This value is closer to the minimum summed organic C
losses (13.5 Tg y− 1) than previously indicated. The Lake Superior
organic carbon budget remains unbalanced but gains and losses are
closer together than they were before. The wide range of estimates of
respiration in the lake (13 to 81 Tg y− 1) is still a large source of
uncertainty in attempting to close Lake Superior's organic carbon
budget. Furthermore, other fluxes such as riverine inputs, bear
additional study and consideration. In addition, input of organic
carbon from aerosols has not yet been estimated for Lake Superior and
might be significant (Duarte et al. 2006).
Perhaps the greatest challenge in using direct measurements such
as these to calculate a value for lake-wide, annual primary production
suitable for budget purposes is spatial and temporal variability. An
ideal study would utilize many deployments across the entire lake for
the entire year, but this would be a highly impractical undertaking in
Lake Superior. We therefore need to consider whether the scaling
approach performed here, using a limited number of samplings, is
adequate. Rate measurements of any kind in the offshore of Lake
Superior during the winter have rarely been reported. The model
developed here for volumetric production as a function of temperature and light was a good descriptor of summer, November and April
conditions at all depths ≥ 5 m (Fig. 2C). The variability of factors that
might affect primary production in this large lake is still incompletely
characterized. Regarding spatial variation at any point in time,
Barbiero and Tuchman (2001, 2004) provide information on lakewide chlorophyll values with depth, and from their data, it is difficult
to discern any stable spatial gradients or differences in the offshore
chlorophyll levels across the lake. Thus, considering the lake as a
single homogeneous system provides at least a provisional estimate,
pending better information on spatial variation in the lake.
The model developed here suggests that light and temperature are
key factors. A muted seasonal variability in chlorophyll levels points to
an overriding importance of incident light in affecting the lake's light
climate. Seasonal variation in incident light driven both by astronomical factors including solar angle and daylight, and by differences in
cloudiness across seasons has been incorporated in the present model.
Regarding temperature, the south shore often has higher temperatures and different stratification patterns than the rest of the lake
early in the summer (Chen et al. 2001). The two sampling sites used
here are well outside of that near-shore zone. The DCM is a potentially
important component of spatial variability. The model developed here
had good power to predict volumetric production with just
temperature and light. The fact that including chlorophyll concentration worsened predictive power likely derives from two factors. First,
the range of chlorophyll contents was not large. Offshore Lake
Superior does not have wide seasonal swings in algal biomass (see
Supplementary Table 1). Second, the DCM depth did not exhibit
relatively high levels of production. A model for production lacking
algal biomass is clearly not a mechanistic model, but the purpose here
was to build a predictive model. The model should not be expected to
operate outside of the range of conditions observed here. The estimate
derived here, based as it is on still limited spatial and temporal
coverage, rests on the finding that volumetric production is well
predicted by light and temperature. It further assumes that the spatial
variability in light and temperature can be overlooked so that they can
be characterized across time with reasonable accuracy.
9
Allowing for rapid cycling of fixed organic carbon into the
dissolved pool was partially responsible for the upward revision of
primary production in this paper. Estimates of the percent of primary
production in different lakes released by algae as DOC vary widely.
There have not been many studies of algal DOC exudation in the Great
Lakes. Laird et al. (1986) found 11 ± 9% (mean ± SD) for a study
performed on epilimnetic Lake Michigan water. Lee and Nalewajko
(1978) reported even lower values (1.2%, Lake Erie and 4.4%, Lake
Ontario). Urban et al. (2005) on the other hand report a value of 35%
for Lake Superior. The value for Lake Superior (22%) reported here is
within the range of relevant reported values.
Based perhaps most of all on its high transparency, Lake Superior
has often been described as “ultraoligotrophic”. Classifying lakes into
trophic categories is imprecise and is generally done only for
convenience. However, the mean water column production of Lake
Superior ( in round numbers ∼300 mg C m− 2 d− 1, Fig. 2D) is actually
at the more productive end of what Wetzel (2001, his Table 15-13)
considered to be representative of “oligotrophic” systems, and it even
extends into the lower end of his “mesotrophic” category. Therefore,
in terms of primary production, Lake Superior should not be
considered an “ultraoligotrophic.” Lakes with lower estimated annual
production than Lake Superior include high latitude lakes and ponds,
high altitude lakes and a mid-latitude hardwater lake (Wetzel 2001,
his Table 15-12). A potential 12-month growing season in Lake
Superior (dependent on ice conditions) contributes to this perhaps
higher than expected level of annual production. The ratio of
maximum to minimum water column production is estimated to be
only two (see Fig. 2D). These considerations strongly suggest that
winter production is hardly negligible in consideration of the annual
organic carbon budget and needs further study. The model suggests
that the added production from the relatively warm and bright mixed
layer in summertime conditions increases lake-wide annual production but not perhaps as much as would be expected.
Although this study suggests a somewhat higher value of primary
production for Lake Superior than previous studies, the overall annual rate
suggested here still remains less than values published for the other
Laurentian Great Lakes. For example, studies on Lake Michigan have
reported 325–350 mg C m− 2 d− 1 (annual, Cotner and Biddanda 2002) or
615–630 mg C m− 2 d− 1 (summertime, Fahnenstiel and Scavia 1987).
These are higher than the 200–300 mg C m− 2 d− 1 reported here for Lake
Superior prior to adjustment for total organic production (Fig. 2D). Studies
on Lake Erie have measured 155–169 g C m− 2 (summertime, Smith et al.,
2005) and 320–370 g C m− 2 (annual, Fitzpatrick et al. 2007). These are
higher than the 93.7 g C m− 2 reported here for Lake Superior for the entire
year. The net primary production for the entire Laurentian Great Lakes
Basin (land plus water) was recently estimated as 346 g C m− 2 for the
entire year (Karim et al. 2008). Alin and Johnson (2007) compiled carbon
cycle information for 41 large lakes of the world and found that primary
production was negatively related to latitude and mean depth and
positively related to solar insulation, mean annual water temperature and
the ratios of watershed–lake area. The primary production of Lake
Superior, at 93.7 g C m− 2 y− 1 (particles) falls within the range of the
global large lakes data (10–1900 g C m− 2 y− 1).
The rate of production calculated here can be directly compared
to certain oceanic studies because of the use of identical
methodology. Karl et al. (1998) reported primary production at
North Pacific Station ALOHA to average 472 mg C m− 2 d− 1
(SD = 125, N = 70). Even the summertime rates observed here for
Lake Superior are among the lowest observed rates at Station
ALOHA. Incorporation of DO14C production potentially raises the
ALOHA value to near 1 g m− 2 d− 1 (Karl et al. 1998) but raises the
Lake Superior values to a lesser extent. The Lake Superior water
column thus has a much lower rate of primary production than the
North Pacific gyre. Production in the temporally more variable
Bermuda Atlantic Time Series (BATS) station fall within a range of
180–810 mg C− 2 d− 1 with annual rates of 100–170 g C− 2 y− 1
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
10
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
(Michaels et al. 1994). The range of values of integrated production
observed here for Lake Superior is similar to that observed during
non-bloom conditions at BATS.
The DCM is a prominent photosynthetic feature of Lake Superior
but its formation and maintenance are not yet well understood.
Chlorophyll concentrations at mid-depth often exceed concentrations
at the surface by factors of two- or three-fold (Fig. 1). The DCM is often
though not always well below the thermocline (this study and
unpublished data). The degree to which the DCM represents light/
shade adaptation of algae, altering chlorophyll content, vs. changes in
biomass is an important consideration, but can only be partially
addressed because reliable estimates of biomass of photosynthesizing
organisms are not available. Nevertheless, depth relations of POC,
which includes heterotrophic and detrital carbon as well as
autotrophic biomass (Supplementary Fig. 1) also indicate mid-depth
maxima, suggesting that the DCM is not entirely due to light/shade
adaptation. Depths of peak POC concentration are shallower than the
DCM and are closer to the thermocline depth.
Fahnenstiel and Glime (1983), in the only other study that
examined depth-dependent production relative to the DCM in Lake
Superior, concluded that the DCM results from in situ production.
They observed a correspondence between the DCM depth and the
depth of maximum primary production. However, observations in the
present study contradict those observations. It is possible that the
factors affecting DCM formation and maintenance changed between
the study of Fahnenstiel and Glime and the present study. Here, the
DCM did not correspond to the stratum of highest productivity, either
in terms of volumetric rates or in terms of productivity to biomass
ratios. In most profiles during stratified conditions, the DCM was
located well below the peak primary production (Table 3 and Fig. 1).
The spatial disjunction of the DCM with the depths of maximum
biomass production and the depth of maximal temperature gradient
argues that, to understand the Lake Superior DCM, we need to look
hard at loss factors including grazing.
Conclusions
Lake Superior is an oligotrophic lake with rates of water column
production ranging from ∼200 to 350 mg C m− 2 d− 1. The phytoplankton community is dominated by small size classes and includes a large
representation of small flagellated taxa. The annual rate of input of
organic carbon to the lake by primary production is estimated here to be
9.73 Tg y− 1. This value is higher than previous estimates that have been
used in constructing the whole-lake annual organic carbon budget,
which helps to close the Lake Superior organic carbon budget.
Role of the funding source
This work is the result of research sponsored by the Minnesota Sea
Grant College Program supported by the NOAA office of Sea Grant,
United States Department of Commerce, under grant No.
NA07OAR4170009. The U.S. Government is authorized to reproduce
and distribute reprints for government purposes, not withstanding
any copyright notation that may appear hereon. This paper is journal
reprint No. JR 561 of the Minnesota Sea Grant College Program. This
material was also based on work supported by the National Science
Foundation, while the author was working at the Foundation.
Acknowledgments
The author thanks Sandra Brovold for a wide range of technical
support and assistance, Kiyoko Yokota for phytoplankton enumeration, James Cotner and Bridget Seegers for discussions, and Mike King,
Doug Ricketts, and the rest of the capable crew of the R/V Blue Heron
for logistical support. James Hood and three anonymous reviewers
provided helpful comments.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jglr.2009.12.007.
References
Alin, S.R., Johnson, T.C., 2007. Carbon cycling in large lakes of the world: a synthesis of
production, burial, and lake-atmosphere exchange estimates. Glob. Biogeochem.
Cycles 21, GB3002.
Anagnostou, E., Sherrell, R., 2008. A MAGIC method for sub-nanomolar orthophosphate
determination in freshwater. Limnol. Oceanogr.: Methods 6, 64–74.
Austin, J.A., Colman, S.M., 2007. Lake Superior summer water temperatures are
increasing more rapidly than regional air temperatures: a positive ice–albedo
feedback. Geophys. Res. Lett. 34.
Austin, J.A., Colman, S.M., 2008. A century of temperature variability in Lake Superior.
Limnol. Oceanogr. 53, 2724–2730.
Baehr, M.M., McManus, J., 2003. The measurement of phosphorus and its spatial and
temporal variability in western Lake Superior. J. Great Lakes Res. 29, 479–487.
Barbiero, R.P., Tuchman, M.L., 2001. Results from the U.S. EPA's biological open water
surveillance program of the Laurentian Great Lakes: II. Deep chlorophyll maxima.
J. Great Lakes Res. 27, 155–166.
Barbiero, R.P., Tuchman, M.L., 2004. The deep chlorophyll maximum in Lake Superior.
J. Great Lakes Res. 30 (Supplement I), 256–268.
Chen, C.S., Zhu, J., Ralph, E.A., Green, S.A., Budd, J., Zhang, F.Y, 2001. Prognostic modeling
studies of the Keweenaw Current in Lake Superior. Part I: formation and evolution.
J. Phys. Oceanogr. 31, 379–395.
Cotner, J., Biddanda, B., Makino, W., Stets, E., 2004. Organic carbon biogeochemistry of
Lake Superior. Aquat. Ecosyst. Health Manag. 7, 451–464.
Cotner, J.B., Biddanda, B.A., 2002. Love handles in aquatic ecosystems: the role of
dissolved organic carbon drawdown, resuspended sediments, and terrigenous
inputs in the carbon balance of Lake Michigan. Ecosystems 5, 431–445.
Duarte, C.M., Dachs, J., Llabrés, M., Alonso-Laita, P., Gasol, J.M., Tovar-Sánches, A.,
Sañudo-Wilhemy, S., Agustí, S., 2006. Aerosol inputs enhance new production in
the subtropical northeast Atlantic. J. Geophys. Res. 111.
Eppley, R.W., 1972. Temperature and phytoplankton growth in the sea. Fish. Bull. 70,
1063–1085.
Fahnenstiel, G.L., Beckmann, C., Lorhrez, S.E., Millie, D.F., Schofield, O.M.E., McCormick,
M.J., 2002. Standard Niskin and Van Dorn bottles inhibit phytoplankton
photosynthesis in Lake Michigan. Internationale Vereinigung für theoretische
und angewandte Limnologie. Verhandlungen 28, 376–380.
Fahnenstiel, G.L., Glime, J.M., 1983. Subsurface chlorophyll maximum and associated
Cyclotella pulse in Lake Superior, USA. Int. Rev. Gesamten Hydrobiol. 68, 605–616.
Fahnenstiel, G.L., Scavia, D., 1987. Dynamics of Lake Michigan phytoplankton: primary
production and growth. Can. J. Fish. Aquat. Sci. 44, 499–508.
Fahnenstiel, G.L., Sicko-Goad, L., Scavia, D., Stoermer, E.F., 1986. Importance of
picoplankton in Lake Superior. Can. J. Fish. Aquat. Sci. 43, 235–240.
Fee, E.J., Shearer, J.A., Debruyn, E.R., Schindler, E.U., 1992. Effects of lake size on
phytoplankton photosynthesis. Can. J. Fish. Aquat. Sci. 49, 2445–2459.
Field, M.P., Sherrell, R.M., 2003. Direct determination of ultra-trace levels of metals in
fresh water using desolvating micronebulization and HR–ICP–MS: application to
Lake Superior waters. J. Anal. Atomic Spectrom. 18, 254–259.
Fitzpatrick, M.A.J., Munawar, M., Leach, J.H., Haffner, G.D., 2007. Factors regulating
primary production and phytoplankton dynamics in western Lake Erie. Fundam.
Appl. Limnol./Archiv. Hydrobiol. 169, 137–152.
Gillooly, J.F., Brown, J.H., West, G.B., Savage, V.M., Charnov, E.L., 2002. Effects of size and
temperature on metabolic rate. Science 293, 2248–2251.
Ivanikova, N.V., Matteson, A.R., Wilhelm, S.W., McKay, R.M.L., Gobler, C.J., Bullerjahn,
G.S., 2008. Picoplanktonic cyanobacteria in Lakes Superior and Erie: phylogenies
of endemic populations and cultured isolates. Internationale Vereinigung für
theoretische und angewandte Limnologie. Verhandlungen 30, 459–465.
Ivanikova, N.V., Popels, L.C., McKay, R.M.L, Bullerjahn, G., 2007. Lake Superior supports
novel clusters of cyanobacterial picoplankton. Appl. Environ. Microbiol. 73, 4055–4065.
Jassby, A.D., Platt, T., 1976. Mathematical formulation of the relationship between
photosynthesis and light for phytoplankton. Limnol. Oceanogr. 21, 540–547.
Karim, A., Veizer, J., Barth, J., 2008. Net ecosystem production in the great lakes basin
and its implications for the North American missing carbon sink: a hydrologic and
stable isotope approach. Glob. Planet. Change 61, 15–27.
Karl, D.M., Hebel, D.V., Björkman, K., Letelier, R.M., 1998. The role of dissolved organic
matter release in the productivity of the oligotrophic North Pacific Ocean. Limnol.
Oceanogr. 43, 1270–1286.
Knap, A., Michaels, A., Close, A., Ducklow, H., Dickson, A, 1996. Protocols for the Joint
Global Ocean Flux Study (JGOFS) Core Measurements. UNESCO, p. 170.
Kumar, S., Sterner, R.W., Finlay, B.J., Brovold, S., 2007. Spatial and temporal variation of
ammonium in Lake Superior. J. Great Lakes Res. 33, 581–591.
Kumar, S., Sterner, R.W., Finlay, J., 2008. Nitrogen and carbon uptake dynamics in Lake
Superior. J. Geophys. Res. – Biogeosci. 113, G04003.
Laird, G.A., Scavia, D., Fahnenstiel, G.L., 1986. Algal organic carbon excretion in Lake
Michigan. J. Great Lakes Res. 12, 136–141.
Lee, K., Nalewajko, C., 1978. Photosynthesis, extracellular release and glucollic acid
uptake by plankton: fractionation studies. Internationale Vereinigung fur theoretische und angewandte Limnologie. Verhandlungen. 20, 257–262.
Michaels, F., Knap, A.H., Dow, R.L., Gundersen, K., Johnson, R.J., Sorensen, J., Close, A.,
Knauer, G.A., Lohrenz, S.E., Asper, V.A., Tuel, M., Bidigare, R., 1994. Seasonal patterns
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007
ARTICLE IN PRESS
R.W. Sterner / Journal of Great Lakes Research xxx (2010) xxx–xxx
of ocean biogeochemistry at the U.S. JGOFS Bermuda Atlantic time-series study site.
Deep-Sea Res. I (41), 1013–1038.
Moll, R.A., Stoermer, E.F., 1982. A hypothesis relating trophic status and subsurface
chlorophyll maxima of lakes. Arch. Hydrobiol. 94, 425–440.
Munawar, M., Munawar, I.F., 1978. Phytoplankton of Lake Superior 1973. J. Great Lakes
Res. 4, 415–442.
Nalewajko, C., Lee, K., Shear, H., 1981. Phosphorus kinetics in Lake Superior: light
intensity and phosphate uptake in algae. Can. J. Fish. Aquat. Sci. 38, 224–323.
Nalewajko, C., Voltolina, D., 1986. Effects of environmental variables on growth rates
and physiological characteristics of Lake Superior phytoplankton. Can. J. Fish.
Aquat. Sci. 43, 1163–1170.
Olson, T.A., Odlaug, T.O., 1966. Limnological observation on Western Lake Superior.
Univ. Mich. Great Lakes Res. Div. Publ. 15, 109–118.
Parkos, W.G., Olson, T.A., Odlaug, T.O., 1969. Water quality studies on the Great Lakes
based on C-14 measurements on primary productivity. Univ. Minn. Water Resour.
Res. Cent. Bull. 17, 121.
11
Putnam, H.D., Olson, T.A., 1961. Studies on the productivity and plankton of Lake
Superior. School of Public Health. University of Minnesota.
Putnam, H.D., Olson, T.A., 1966. Primary productivity at a fixed station in Western Lake
Superior. Proc. 9th Conf. Great Lakes Res. Int. Assoc: Great Lakes Res., pp. 115–128.
Smith, R.E.H., Hiriart-Baer, V.P., S.N., H., Guildford, S.J., Charlton, M.N., 2005. Planktonic
primary production in the offshore waters of Dreissenid-infested Lake Erie in 1997.
J. Great Lakes Res. 31, 50–62.
Sterner, R.W., Smutka, T.M., McKay, R.M.L., Xiaoming, Q., Brown, E.T., Sherrell, R.M.,
2004. Phosphorus and trace metal limitation of algae and bacteria in Lake Superior.
Limnol. Oceanogr. 49, 495–507.
Urban, N.R., Auer, M.T., Green, S.A., Lu, X., Apul, S., Powell, K.D., Bub, L., 2005. Carbon
cycling in Lake Superior. J. Geophys. Res. 110, C06S90.
Vollenweider, R.A., Munawar, M., Stadelmann, P., 1974. A comparative review of
phytoplankton and primary production in the Laurentian Great Lakes. J. Fish. Res.
Board of Can. 31, 739–762.
Wetzel, R.G., 2001. Limnology. Lake and River Ecosystems. Academic Press, San Diego.
Please cite this article as: Sterner, R.W., In situ-measured primary production in Lake Superior, J. Great Lake Res. (2010), doi:10.1016/j.
jglr.2009.12.007